Maternal behaviours may explain riffle‐scale variations in some stream insect populations

Abstract

Summary Species composition and abundance of larval insects can vary significantly between similar channel units (riffles) on the same stream as well as between similar streams. The source of this variation is largely unexplained, but important because such spatial variations can influence population and community dynamics and the design of sampling programmes. The maternal and oviposition behaviours of some aquatic insects are specialised, such that eggs are laid in particular locations, and this could influence the spatial distribution of eggs and larvae within a stream. In a relatively short length of stream (1 km), we tested whether the number of oviposition sites influenced the number of larvae of different instars in whole riffles that ranged in area from 80 to 460 m2. We examined two species, a mayfly Baetis rhodani and a caddisfly Rhyacophila dorsalis, with larvae that are riffle specialists and adults that oviposit exclusively on emergent rocks in riffles. For comparison, we also examined a control species, blackflies Simulium spp., with larvae that are also riffle specialists, but adults that oviposit on trailing vegetation. As an alternative hypothesis, we tested whether the numbers of larvae per riffle were associated with habitat availability, as measured by riffle area. Numbers of early‐instar B. rhodani and R. dorsalis were positively associated with the numbers of oviposition sites per riffle, but not riffle areas. In contrast, numbers of early‐instar Simulium per riffle (the control species) were not associated with either predictor variable. For later instars, riffle area was a good predictor of larval numbers per riffle for all species, even though the number of oviposition sites still explained much variation for B. rhodani and R. dorsalis. This ontogenetic shift in distribution patterns suggests an adjustment of densities during larval life, and the evidence points to density‐dependent mortality of early‐instar larvae. Contrary to the prevailing view that riffle‐to‐riffle variations in the numbers of aquatic insect larvae can be explained by larval habitat, this study demonstrates a strong influence of adult behaviour, and oviposition site availability can explain significant amounts of this variation for some species. These results challenge traditional views about aquatic insect populations, which may have more complex and interesting dynamics than considered previously.